4 ms·
With a different ranking formula it's actually quite promising. I created a simple simulation and some plots: https://github.com/fdietze/downvote-scoring https:
by manx 5y ago
With a different ranking formula it's actually quite promising. I created a simple simulation and some plots: https://github.com/fdietze/downvote-scoring https://github.com/fdietze/downvote-scoring
- jokoon 5y agoI don't understand this simulation, did you have real users who could only downvote?
- manx 5y agoNo real users. It's a simplified simulation of a frontpage and how random users interact with it. Have a look at the code: https://github.com/fdietze/downvote-scoring/blob/master/scoring-simulation.scala#L104 https://github.com/fdietze/downvote-scoring/blob/master/scor... I started to write another simulation with the goal to be more realistic, i.e. backed by real statistics from HN, like "how many votes arrive on each rank of the frontpage", or "what is the distribution of submission arrival".